import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import plotly.express as px
import warnings
warnings.filterwarnings('ignore')

#!pip install plotly_express==0.4.0

df = pd.read_csv("C:/Users/Zehra Salmani/Desktop/extras/python/Instagram data.csv",encoding="ISO-8859-1")

df.head()

	Impressions	From Home	From Hashtags	From Explore	From Other	Saves	Comments	Shares	Likes	Profile Visits	Follows	Caption	Hashtags
0	3920	2586	1028	619	56	98	9	5	162	35	2	Here are some of the most important data visua...	#finance #money #business #investing #investme...
1	5394	2727	1838	1174	78	194	7	14	224	48	10	Here are some of the best data science project...	#healthcare #health #covid #data #datascience ...
2	4021	2085	1188	0	533	41	11	1	131	62	12	Learn how to train a machine learning model an...	#data #datascience #dataanalysis #dataanalytic...
3	4528	2700	621	932	73	172	10	7	213	23	8	Here’s how you can write a Python program to d...	#python #pythonprogramming #pythonprojects #py...
4	2518	1704	255	279	37	96	5	4	123	8	0	Plotting annotations while visualizing your da...	#datavisualization #datascience #data #dataana...

df.shape

(119, 13)

df.dtypes

Impressions        int64
From Home          int64
From Hashtags      int64
From Explore       int64
From Other         int64
Saves              int64
Comments           int64
Shares             int64
Likes              int64
Profile Visits     int64
Follows            int64
Caption           object
Hashtags          object
dtype: object

df.isnull().sum()

Impressions       0
From Home         0
From Hashtags     0
From Explore      0
From Other        0
Saves             0
Comments          0
Shares            0
Likes             0
Profile Visits    0
Follows           0
Caption           0
Hashtags          0
dtype: int64

plt.figure(figsize=(8,6))
plt.title("Impressions from Home")
sns.distplot(df['From Home'])
plt.show()

plt.figure(figsize=(8,6))
plt.title("Impressions from Hashtags")
sns.distplot(df['From Hashtags'])
plt.show()

plt.figure(figsize=(8,6))
plt.title("Impressions from Explore")
sns.distplot(df['From Explore'])
plt.show()

plt.figure(figsize=(8,6))
plt.title("Impressions from Other")
sns.distplot(df['From Other'])
plt.show()

home = df['From Home'].sum()
hashtags = df['From Hashtags'].sum()
explore = df['From Explore'].sum()
others = df['From Other'].sum()

labels = ['From Home','From Hashtags','From Explore','From Other']
values = [home,hashtags,explore,others]

fig = px.pie(df,names = labels,values = values,title='Impressions from Various Sources',hole=0.5)
fig.show()

from PIL import Image
from wordcloud import WordCloud, STOPWORDS, ImageColorGenerator

#!pip install wordcloud

#for all captions
text = ' '.join(i for i in df.Caption)
print("There are {} words in combination of all captions ".format(len(text)))

There are 23021 words in combination of all captions 

stopwords = set(STOPWORDS)
#generate a word cloud image

wordcloud = WordCloud(mode = "RGBA",width = 800,height=500,max_font_size=100,stopwords=stopwords,max_words=80,
                      background_color=None).generate(text)
plt.imshow(wordcloud,interpolation = 'quadric')
plt.axis("off")
plt.show()

#for all hashtags
text1 = ' '.join(i for i in df.Hashtags)
print("There are {} words in combination of all hashtags ".format(len(text)))

There are 23021 words in combination of all hashtags 

stopwords = set(STOPWORDS)
#generate a word cloud image

wordcloud = WordCloud(mode = "RGBA",width = 500,height=300,max_font_size=100,stopwords=stopwords,max_words=50,
                      background_color=None).generate(text1)
plt.imshow(wordcloud,interpolation = 'quadric')
plt.axis("off")
plt.show()

fig = px.scatter(data_frame = df,x='Impressions',y='Likes',size='Likes',trendline = 'ols',
                title = 'Relationship between Impressions and Likes')
fig.show()

Theres a linear relationship between impressions and likes, likes increase as impressions increase

fig = px.scatter(data_frame = df,x='Impressions',y='Comments',size='Comments',trendline = 'ols',
                title = 'Relationship between Impressions and Comments')
fig.show()

Theres no relation between impressions and comments

fig = px.scatter(data_frame = df,x='Impressions',y='Shares',size='Shares',trendline = 'ols',
                title = 'Relationship between Impressions and Shares')
fig.show()

A more number of shares will result in a higher reach, but shares don’t affect the reach of a post as much as likes do

fig = px.scatter(data_frame = df,x='Impressions',y='Saves',size='Saves',trendline = 'ols',
                title = 'Relationship between Impressions and Saves')
fig.show()

Theres a linear relationship between Impressions and Saves

fig = px.scatter(data_frame = df,x='Impressions',y='Profile Visits',size='Profile Visits',trendline = 'ols',
                title = 'Relationship between Impressions and Profile Visits')
fig.show()

df.corr()

	Impressions	From Home	From Hashtags	From Explore	From Other	Saves	Comments	Shares	Likes	Profile Visits	Follows
Impressions	1.000000	0.844698	0.560760	0.893607	0.592960	0.779231	-0.028524	0.634675	0.849835	0.760981	0.889363
From Home	0.844698	1.000000	0.177516	0.800573	0.555666	0.768817	0.012716	0.674985	0.698330	0.531076	0.672675
From Hashtags	0.560760	0.177516	1.000000	0.190453	0.229623	0.305929	0.161439	0.219511	0.662124	0.691345	0.555485
From Explore	0.893607	0.800573	0.190453	1.000000	0.495685	0.747803	-0.158565	0.615731	0.653699	0.531850	0.796019
From Other	0.592960	0.555666	0.229623	0.495685	1.000000	0.331907	-0.108703	0.156834	0.393510	0.633080	0.546737
Saves	0.779231	0.768817	0.305929	0.747803	0.331907	1.000000	-0.026912	0.860324	0.845643	0.360628	0.628461
Comments	-0.028524	0.012716	0.161439	-0.158565	-0.108703	-0.026912	1.000000	0.016933	0.123586	0.096714	-0.060631
Shares	0.634675	0.674985	0.219511	0.615731	0.156834	0.860324	0.016933	1.000000	0.707794	0.245361	0.493070
Likes	0.849835	0.698330	0.662124	0.653699	0.393510	0.845643	0.123586	0.707794	1.000000	0.626107	0.746333
Profile Visits	0.760981	0.531076	0.691345	0.531850	0.633080	0.360628	0.096714	0.245361	0.626107	1.000000	0.853152
Follows	0.889363	0.672675	0.555485	0.796019	0.546737	0.628461	-0.060631	0.493070	0.746333	0.853152	1.000000

a = df.corr()
a['Impressions'].sort_values(ascending = False)

Impressions       1.000000
From Explore      0.893607
Follows           0.889363
Likes             0.849835
From Home         0.844698
Saves             0.779231
Profile Visits    0.760981
Shares            0.634675
From Other        0.592960
From Hashtags     0.560760
Comments         -0.028524
Name: Impressions, dtype: float64

So we can say that more likes and saves will help you get more reach on Instagram. The higher number of shares will also help you get more reach, but a low number of shares will not affect your reach either.

In Instagram, conversation rate means how many followers you are getting from the number of profile visits from a post.

conversation_rate = (df['Follows'].sum()/df['Profile Visits'].sum())*100
print(conversation_rate)

41.00265604249668

fig = px.scatter(data_frame = df,y='Follows',x='Profile Visits',size='Profile Visits',trendline = 'ols',
                title = 'Relationship between  Profile Visits and Follows ')
fig.show()

df.columns

Index(['Impressions', 'From Home', 'From Hashtags', 'From Explore',
       'From Other', 'Saves', 'Comments', 'Shares', 'Likes', 'Profile Visits',
       'Follows', 'Caption', 'Hashtags'],
      dtype='object')

Instagram Reach Prediction Model

x = np.array(df[['Saves', 'Comments', 'Shares', 'Likes', 'Profile Visits',
       'Follows']])
y = np.array(df['Impressions'])

from sklearn.model_selection import train_test_split
from sklearn.linear_model import PassiveAggressiveRegressor

x_train, x_test, y_train, y_test = train_test_split(x, y, random_state=42, test_size = 0.2)

model = PassiveAggressiveRegressor()

model.fit(x_train,y_train)

PassiveAggressiveRegressor()

model.predict(x_test)

array([ 6230.79478639,  9569.85722763,  3811.04600369,  7668.73069153,
        5797.44215179, 29597.41308771,  3969.06578757,  2977.03593739,
       13877.74412764,  4991.34990147,  6871.64146821,  9898.47750504,
        4843.31138824,  3601.15659496,  5408.08030377,  4666.33705043,
        4008.83479828,  5463.62461192,  5059.12675504,  6043.55445996,
       11514.06016197, 21789.62933606,  5797.44215179,  7293.3771316 ])

model.score(x_test,y_test)

0.8883558388190361

# Features = [['Saves', 'Comments', 'Shares', 'Likes', 'Profile Visits','Follows']]
features = np.array([[233.0,4.0,9.0,282.0,165.0, 54.0]])
model.predict(features)

array([13533.46854936])